Vehicular Mobility Simulation for VANETs∗
Marco Fiore
Politecnico di Torino
Corso Duca degli Abruzzi 24
10129 Torino, Italy
Jerome Harri, Fethi Filali, Christian Bonnet
Institut Eurecom †
Department of Mobile Communications
06904 Sophia-Antipolis, France
{haerri,filali,bonnet}@eurecom.fr
Abstract
During the last few years, continuous progresses in wireless
communications have opened new research fields in computer
networking, aimed at extending data networks connectivity to
environments where wired solutions are impracticable. Among
these, vehicular traffic is attracting a growing attention from
both academia and industry, due to the amount and impor-
tance of the related applications, ranging from road safety to
traffic control, up to mobile entertainment. Vehicular Ad-hoc
Networks (VANETs) are self-organized networks built up from
moving vehicles, and are part of the broader class of Mobile
Ad-hoc Networks (MANETs). Because of their peculiar char-
acteristics, VANETs require the definition of specific network-
ing techniques, whose feasibility and performance are usually
tested by means of simulation. One of the main challenges
posed by VANETs simulations is the faithful characterization of
vehicular mobility at both macroscopic and microscopic levels,
leading to realistic non-uniform distributions of cars and ve-
locity, and unique connectivity dynamics. In this paper we first
present and describe VanetMobiSim, a freely available gener-
ator of realistic vehicular movement traces for networks simu-
lators.Then, VanetMobiSim is validated by illustrating how the
interaction between featured macro- and micro-mobility is able
to reproduce typical phenomena of vehicular traffic.
1. Introduction
Vehicular Ad-hoc Networks (VANETs) represent a rapidly
emerging, particularly challenging class of Mobile Ad Hoc
Networks (MANETs). VANETs are distributed, self-
organizing communication networks built up from traveling
∗This work has been supported partially by the European Community
through the NoE NEWCOM, and partially by the Institut Eurecom and the
Politecnico di Torino.†Institut Eurecom’s research is partially supported by its industrial
members: BMW Group Research & Technology - BMW Group Com-
pany, Bouygues Telecom, Cisco Systems, France Telecom, Hitachi Eu-
rope, SFR, Sharp, STMicroelectronics, Swisscom, Thales.
vehicles, and are thus characterized by very high speed and
limited degrees of freedom in nodes movement patterns. Such
particular features often make standard networking protocols
inefficient or unusable in VANETs, and this, combined with
the huge impact that the deployment of VANET technologies
could have on the automotive market, explains the growing ef-
fort in the development of communication protocols which are
specific to vehicular networks.
Whereas it is crucial to test and evaluate protocol imple-
mentations in real testbed environments, logistic difficulties,
economic issues and technology limitations make simulation
the mean of choice in the validation of networking protocols
for VANETs, and a widely adopted first step in development of
real world technologies. A critical aspect in a simulation study
of VANETs, is the need for a mobility model which reflects, as
close as possible, the real behavior of vehicular traffic. When
dealing with vehicular mobility modeling, we distinguish be-
tween macro-mobility and micro-mobility descriptions.
For macro-mobility, we intend all the macroscopic aspects
which influence vehicular traffic: the road topology, con-
straining cars movement, the per-road characterization defin-
ing speed limits, number of lanes, overtaking and safety rules
over each street of the aforementioned topology, or the traffic
signs description establishing the intersections crossing rules.
Micro-mobility instead refers to the drivers’ individual be-
havior, when interacting with other drivers or with the road
infrastructure: traveling speed in different traffic conditions;
acceleration, deceleration and overtaking criteria, behavior in
presence of road intersections and traffic signs, general driving
attitude related to driver’s age, sex or mood, etc.
It would be desirable for a trustworthy VANETs simulation
that both macro-mobility and micro-mobility descriptions be
jointly considered in modeling vehicular movements. Indeed,
many non-specific mobility models employed in VANETs sim-
ulations ignore these guidelines, and thus fail to reproduce pe-
culiar aspects of vehicular motion, such as car acceleration and
deceleration in presence of nearby vehicles, queuing at road
intersections, clustering caused by semaphores, vehicular con-
gestion and traffic jams.
Proceedings of the 40th Annual Simulation Symposium (ANSS'07)0-7695-2814-7/07 $20.00 © 2007
Authorized licensed use limited to: IEEE Xplore. Downloaded on March 25, 2009 at 05:53 from IEEE Xplore. Restrictions apply.
In this paper, we introduce VanetMobiSim [3], a freely dis-
tributed, open source vehicular mobility generator based on
the CanuMobiSim architecture [4] and designed for integra-
tion with telecommunication network simulators. VanetMo-
biSim can produce detailed vehicular movement traces em-
ploying different macro- and micro-mobility models and taking
into account the interaction of the two, and can simulate differ-
ent traffic conditions through fully customizable scenarios. We
validate the mobility patterns generated by VanetMobiSim by
recreating distinctive vehicular mobility effects, such as speed
decay with increasing car density, non-uniform distribution of
vehicles in urban areas, and shock waves due to stop-and-go
perturbations.
Due to space limitations, we could not include a detailed
related work on the current state-of-the-art of vehicular mobil-
ity modeling for VANETs simulations. We refer the interested
reader to [1], which surveys and classifies major models avail-
able to the community. Moreover, an extended version of the
present work is also available [2].
The rest of the paper is organized as follows. A detailed de-
scription of the features of VanetMobiSim is given in Section 2.
Section 3 presents some tests validating the movement traces
produced by VanetMobiSim in specific scenarios. Finally, in
Section 4, we draw some conclusions and discuss future work.
2. VanetMobiSim
VanetMobiSim is an extension to CanuMobiSim [4], a
generic user mobility simulator. CanuMobiSim is a platform-
and simulator-independent software, coded in Java and produc-
ing mobility traces for different network simulators, including
ns-2 [5], QualNet [6] and GloMoSim [7]. It provides an eas-
ily extensible mobility architecture, but, due to its general pur-
pose nature, suffers from a reduced level of detail in specific
scenarios. VanetMobiSim is therefore aimed at extending the
vehicular mobility support of CanuMobiSim to a higher degree
of realism. In this section, we outline the structure and char-
acteristics of VanetMobiSim and detail the resulting vehicular
mobility support.
2.1. Macro-mobility Features
When considering macro-mobility we not only take into ac-
count the road topology, but also the road structure (unidirec-
tional or bidirectional, single- or multi-lane), the road charac-
teristics (speed limits, vehicle-class based restrictions) and the
presence of traffic signs (stop signs, traffic lights, etc.). More-
over, the concept of macro-mobility also includes the effects of
the presence of points of interests, which influence movement
patterns of vehicles on the road topology. All these different
aspects of macro-mobility are discussed in details in the re-
mainder of this section.
2.1.1. Road topology definition
The selection of the road topology is a key factor to obtain
realistic results when simulating vehicular movements. Indeed,
the length of the streets, the frequency of intersections, the den-
sity of buildings can greatly affect important mobility metrics
such as the minimum, maximum and average speed of cars, or
their density over the simulated map. VanetMobiSim allows to
define the road topology in the following ways, the first two
being already implemented in CanuMobiSim:
• User-defined graph: the road topology is specified by list-
ing the vertices of the graph and their interconnecting
edges.
• GDF map: the road topology is imported from a Geo-
graphical Data File (GDF) [8]. Unfortunately, most GDF
file libraries are not freely accessible.
• TIGER map: the road topology is extracted from a map
obtained form the TIGER database [9]. The level of detail
of the maps in the TIGER database is not as high as that
provided by the GDF standard, but this database is open
and contains digital descriptions of wide urban and rural
areas of all districts of the United States.
• Clustered Voronoi graph: the road topology is randomly
generated by creating a Voronoi tessellation on a set of
non-uniformly distributed points. This approach is simi-
lar to that proposed in [10], but we also consider the pres-
ence of areas with different road densities which we refer
to as clusters.
In all these cases, the road topology is implemented as
a graph over whose edges the movement of vehicles is con-
strained. Examples of different VanetMobiSim topologies are
illustrated in Figure 1.
2.1.2. Road topology characterization
As stated before, the concept of vehicular macro-mobility
is not limited to motion constraints obtained from graph-based
mobility, but also includes all aspects related to the road struc-
ture characterization, such as directional traffic flows or multi-
ple lanes, speed constraints or intersection crossing rules. None
of these aspects is present in CanuMobiSim, thus the following
enhancements are introduced by VanetMobiSim:
• introduction of roads with multiple lanes in each direction
• physical separation of opposite traffic flows on each road.
• definition of independent speed limits on each road of the
topology
Proceedings of the 40th Annual Simulation Symposium (ANSS'07)0-7695-2814-7/07 $20.00 © 2007
Authorized licensed use limited to: IEEE Xplore. Downloaded on March 25, 2009 at 05:53 from IEEE Xplore. Restrictions apply.
(a) User-defined topology (b) GDF map topology (c) TIGER map topology (d) Clustered Voronoi
Figure 1. Road topologies examples
• implementation of traffic signs at each road intersection.
By default, intersections are fully regulated by stop signs,
forcing vehicles to stop and wait for free road before
crossing. Alternatively, it is possible to regulate traffic
at intersections by means of traffic lights, whose tempo-
rization is customizable.
2.1.3. Vehicular movement patterns selection
Vehicular traffic schemes in urban scenarios are far from
being random. Indeed, cars tend to move between points of
interests, which are often common to many drivers and can
change in time (e.g., offices may be strong attraction points,
but mainly during the first part of the morning). Accordingly,
VanetMobiSim exploits CanuMobiSim capability of building
up movement patterns from the cooperation of a trip genera-
tion module, which defines the sets of points of interest, and
a path computation module, whose task is to compute the best
path between those points.
Two choices are given for the trip generation module. The
first is a random trip, as the start and stop points of movement
patterns are randomly selected among the vertices of the graph
representing the road topology. The second is an activity se-
quences generation, in which a set of start and stop points are
explicitly provided in the road topology description, and cars
are forced to move among them.
Independently from the trip generation method employed,
the path computation, i.e. the selection of the best sequence
of edges to reach the selected destination, can be performed in
three ways. The first method selects the shortest path to desti-
nation, running a Dijkstra’s algorithm with edges cost inversely
proportional to their length. The second method does not only
considers the length of the path, but also the traffic congestion
level, by weighting the cost of traversing an edge also on the
number of cars traveling on it, thus modeling the real world
tendency of drivers to avoid crowded paths. The last method,
which is not present in the original CanuMobiSim, extends the
other two, by also accounting for the road speed limit when
calculating the cost of an edge, in a way that fastest routes are
preferred.
The combination of trip generation and path computation
methods offers a wide range of possibilities, when the defini-
tion of vehicular movement paths is a factor of interest in the
mobility simulation.
2.2. Micro-Mobility Features
The concept of vehicular micro-mobility includes all as-
pects related to an individual car’s speed and acceleration mod-
eling. The micro-mobility description plays the main role in
the generation of realistic vehicular movements, as it is respon-
sible for effects such as smooth speed variation, cars queues,
traffic jams and overtakings.
Three broad classes of micro-mobility models, featuring
an increasing degree of detail, can be identified depending on
whether the individual speed of vehicles is computed i) in a de-
terministic way, ii) as a function of nearby vehicles behavior in
a single lane scenario, or iii) as a function of nearby vehicles
behavior in a multi-flow interaction (i.e., urban) scenario.
CanuMobiSim provides implementations for models be-
longing to the first two classes. The Graph-Based Mobility
Model (GBMM) [11], the Constant Speed Motion (CSM) [4]
and the Smooth Motion Model (SMM) [12] fall into the first
category, as the speed of each vehicle is determined on the
basis of the local state of each car and any external effect is
ignored. They all constrain a random movement of nodes on
a graph, possibly including pauses at intersections (CSM) or
smooth speed changes when reaching or leaving a destination
(SSM). The movement is random in a sense that vehicles select
one destination and move towards it along a shortest-length
path, ignoring (and thus possibly overlapping) other vehicles
during the motion. While these models may work for isolated
cars, they fail to reproduce realistic movements of groups of
vehicles.
The Fluid Traffic Model (FTM) [13] and Intelligent Driver
Model (IDM) [14] are instead part of the second class, as they
account for the presence of nearby vehicles when calculating
the speed of a car. These models describe car mobility on sin-
Proceedings of the 40th Annual Simulation Symposium (ANSS'07)0-7695-2814-7/07 $20.00 © 2007
Authorized licensed use limited to: IEEE Xplore. Downloaded on March 25, 2009 at 05:53 from IEEE Xplore. Restrictions apply.
gle lanes, but do not consider the case in which multiple vehic-
ular flows have to interact, as in presence of intersections.
The FTM describes the speed as a monotonically decreas-
ing function of the vehicular density, forcing a lower bound on
speed when the traffic congestion reaches a critical state, by
means of the following equation
s = max
ů
smin, smax
Ń
1 − k
kjam
űÿ
where s is the output speed, smin and smax are the minimum
and maximum speed respectively, kjam is the vehicular density
for which a traffic jam is detected, and k is the current vehicular
density of the road the node, whose speed is being computed,
is moving on. This last parameter is given by k = n/l, where
n is the number of cars on the road and l is the length of the
road segment itself.
On the other hand, the IDM characterizes drivers behavior
depending on their front vehicle, thus falling into the so-called
car following models category. The instantaneous acceleration
of a vehicle is computed according to the following equations
dv
dt= a
"
1 −Ń
v
v0
ű 4
−Ń
s∗
s
ű 2#
s∗ = s0 +
Ń
vT +v∆v
2√
ab
ű
In the left hand Equation, v is the current speed of the vehicle,
v0 is the desired velocity, s is the distance from preceding ve-
hicle and s∗ is the so called desired dynamical distance. This
last parameter is computed as shown in the right hand equation,
and is a function of the minimum bumper-to-bumper distance
s0, the minimum safe time headway T , the speed difference
with respect to front vehicle velocity ∆v, and the maximum
acceleration and deceleration values a and b.
VanetMobiSim adds two original microscopic mobility
models, both of which account for the interaction of multiple
converging flows, by acting consistently with the road infras-
tructure, and thus fall into the third category mentioned above.
These models extend the IDM description, which is the most
realistic among those present in CanuMobiSim, in order to
include the management of intersections regulated by traffic
signs and of roads with multiple lanes.
The first new micro-mobility model is referred to as Intel-
ligent Driver Model with Intersection Management (IDM-IM).
It adds intersection handling capabilities to the behavior of ve-
hicles driven by the IDM. In particular, IDM-IM models two
different intersection scenarios: a crossroad regulated by stop
signs, or a road junction ruled by traffic lights. In both cases,
IDM-IM only acts on the first vehicle on each road, as IDM
automatically adapts the behavior of cars following the leading
one. Every time a vehicle finds no intermediate car between
itself and an intersection regulated by stop signs, the following
parameters are used by IDM-IM
(
s = σ − S
∆v = v
where σ is the current distance to the intersection and S is a
safety margin, accounting for the gap between the center of the
intersection and the point the car would actually stop at. Once
a car is halted at a stop sign, it is informed by the macroscopic
level description of the number of cars already waiting to cross
the intersection from any of the incoming roads. If there are no
other cars, the vehicle may pass. Otherwise, it has to wait for
its turn in a first-arrived-first-passed and right hand rule policy.
When a vehicle is heading towards a traffic light intersec-
tion, it is informed by the macroscopic description about the
state of the semaphore. If the color is green, passage is granted
and the car maintains its current speed through the intersection.
If the color is red, crossing is denied and the car is forced to
decelerate and stop at the road junction by using the modified
IDM parameters similarly to a stop sign.
It may also be stressed out that vehicles behavior can dy-
namically vary in presence of traffic lights, according to red-
to-green and green-to-red switches. In the former case, a car
currently decelerating to stop at a red light will accelerate again
if the semaphore turns green before it has completely halted In
the latter case, a vehicle keeping its pace towards a green light
will try to stop if the light becomes red before it has passed
through the intersection. A minimum breaking distance s is
evaluated by means of simple kinematic formulae as
s = v t − κb
2t2 = v
ş v
κb
ť
− κb
2
ş v
κb
ť 2
=v2
2κb
which describes the space needed to come to a full stop as a
function of the current speed of the vehicle, v, the time t and the
the deceleration value, κb. The last parameter represents the
maximum safe deceleration, i.e., the IDM comfortable braking
value b scaled by a factor κ ≥ 1. Upon computation of s, if
the vehicle finds that it is not possible to stop before the inter-
section, even braking as hard as possible, i.e., if s > σ − S,
then it crosses the intersection at its current speed. Otherwise,
it stops by applying a strong enough deceleration. This repro-
duces a real world situation, since, when a traffic light switches
to red, drivers only stop if safety braking conditions can be re-
spected. Examples of driving behaviors in presence of a dy-
namic semaphore are presented in Figure 2.
The second model we introduce is named Intelligent Driver
Model with Lane Changes (IDM-LC), and extends the IDM-
IM model with the possibility for vehicles to change lane
and overtake each others, taking advantage of the multi-lane
capability of the macro-mobility description detailed in Sec-
tion 2.1.2. Two issues are raised by the introduction of multi-
ple lanes: the first is the separation of traffic flows on different
lanes of the same road, while the second is the overtakings
model itself.
Proceedings of the 40th Annual Simulation Symposium (ANSS'07)0-7695-2814-7/07 $20.00 © 2007
Authorized licensed use limited to: IEEE Xplore. Downloaded on March 25, 2009 at 05:53 from IEEE Xplore. Restrictions apply.
0
5
10
15
20
25
30
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Sp
eed
(m
/s)
Time (s)
light turns green after vehicle stoplight turns green before vehicle stop
0
5
10
15
20
25
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Sp
ace
(m x
10
0)
Time (s)
light turns green after vehicle stoplight turns green before vehicle stop
0
5
10
15
20
25
30
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Sp
eed
(m
/s)
Time (s)
light turns red at 40mlight turns red at 100m
0
5
10
15
20
25
30
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Sp
eed
(m
/s)
Time (s)
light turns red at 200mlight turns red at 400m
0
5
10
15
20
25
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Sp
ace
(m x
10
0)
Time (s)
light turns red at 40mlight turns red at 100mlight turns red at 200mlight turns red at 400m
Figure 2. Traffic light red-to-green (left) and green-to-red (right) scenarios. A vehicle, driven by the IDM-IM
setup in Table 1, starts its movement from zero speed, and travels towards a red (green) traffic light, which
then turns green (red) before the car reaches it. The upper figure shows the evolution of speed in time,
while the lower one depicts the car movement on the road versus time. Different curves represent different
driver behaviors, depending on his/her distance from the traffic light when the color switch happens.
As far as the first problem is concerned, vehicular flows on
parallel lanes of the same road are separated by forcing the
car following model to only consider vehicles traveling on the
same lane. However, as the number of lanes can vary from one
road to another, a vehicle approaching a crossroad will receive
from the macro-mobility description the information about the
structure of the road it is going to move to. It can then adopt
one of the following behaviors:
• if the lane the vehicle is currently moving on is also
present in the next road on its path, then it moves through
the intersection and keeps traveling on the same lane in
the next street;
• if the lane currently used by the vehicle does not exist in
the next road, then it tries to merge to its right as it ap-
proaches the junction. If it cannot do it, e.g. because the
lane to its right is very crowded, it stops at the intersection
and waits until a spot becomes available.
On the overtaking model itself, the MOBIL model [15] is
employed, mainly due to its implicit compatibility with the
IDM. This model adopts a game theoretical approach to ad-
dress the lane changing problem, allowing a vehicle to move to
a different lane if the lane change minimizes the overall brak-
ing of vehicles. Such requirement is fulfilled when the two
conditions
al − a ± abias > pş
acur + anew − alcur − al
new
ť
+ athr
alnew > −asafe
are verified. The model allows a vehicle to move to lane l if
the first inequality is verified, that is, if, in terms of accelera-
tion, the advantage of the driver who changes its lane al − a, is
greater than the disadvantages of the following cars in the cur-
rent (acur − alcur) and in the candidate (anew − al
new) lanes.
The MOBIL model also consider a politeness factor p, which
scales the right hand term, in a way that, for values of p to-
wards (or above) one, a polite behavior towards other drivers is
maintained, while, as p moves to (or below) zero, the driver can
become selfish or even malicious. The threshold acceleration
athr introduces a minimum acceleration advantage to allow a
lane change, in order to avoid lane hopping in border cases.
The bias term abias is instead added to favor movements to
one side: in our case, this bias value is added to the advantage
computed for movements to the right and subtracted for move-
ments to the left, thus reproducing the real world tendency of
drivers to stay on their right on a multi-lane road. Finally, in
any case, the safety condition expressed by the right hand side
equation above must be verified for the lane change to occur,
meaning that the new back vehicle does not have to brake too
hard (its deceleration must be over the safe value asafe) as a
consequence of the lane change.
3. VanetMobiSim Validation
Several tests were run on the vehicular movement traces
produced by CanuMobiSim and VanetMobiSim, in order to
verify that the overall mobility description provided by these
tools is able to model vehicular traffic with a sufficient level of
realism. This also gives us the possibility to comment on the
different outputs obtained with various microscopic mobility
models implemented by CanuMobiSim and by VanetMobiSim.
First, different micro-mobility models are tested on a user-
defined graph representing a square city section of 1500 m side.
The urban topology employed in those tests is shown in Fig-
ure 3, where, unless specified differently, all roads have a sin-
Proceedings of the 40th Annual Simulation Symposium (ANSS'07)0-7695-2814-7/07 $20.00 © 2007
Authorized licensed use limited to: IEEE Xplore. Downloaded on March 25, 2009 at 05:53 from IEEE Xplore. Restrictions apply.
������
������
������
������
������
������
������
������
������
������
������
������
������
������
������
������
������
������
������������
������
������
������
������
������
������
A BC
D������
������
��������
��������
1−q
1−p
pq
Figure 3. City section map and activity chain
0
2
4
6
8
10
12
14
16
18
20
10 15 20 25 30 35 40 45 50
Ave
rage
Spe
ed (
m/s
)
Vehicular density (vehicle/km)
RWPCSM
FLUIDIDM
IDM-IM stopsIDM-IM lights
IDM-LC
Figure 4. Average speed versus car density
gle lane, and a speed limit of 15 m/s (54 km/h), except for the
roads represented with thicker lines, which allow a maximum
speed of 20 m/s (72 km/h). Vehicles travel between entry/exit
points at borders, identified with circles and squares, crossing
the city section according to the fastest path to their destination.
The trips generation scheme is activity-based (see Sec-
tion 2.1.3), and the relative transition probability matrix de-
scribes a simple activity chain, depicted in Figure 3. There,
the states denote the class of the selected destination: a round
for the entry/exit points of high-speed roads, a square for the
entry/exit points of normal-speed roads, as also shown in Fig-
ure 3.
The number of cars traveling at the same time within the
city section ranges from 100 to 500, reproducing light (10 vehi-
cles/km) to heavy (50 vehicles/km) traffic conditions. For each
test, a single simulation was run, with statistics recorded for
3600 s, after a transient of 900 s. When computing 95% con-
fidence intervals for mean values collected averaging in time
and on the whole road topology, the error margin was found
to be within 0.5% from the mean. However, we point out that
vehicular traffic in presence of driver finite reaction times and
200 400
600 800
1000 1200
1400 0
200
400
600
800
1000
1200
1400
0
3
6
9
12
Average Vehicular Density(vehicles / 625m
2)
X (m)
Y (m)
Average Vehicular Density(vehicles / 625m
2)
Figure 5. Vehicular density: FTM
200 400
600 800
1000 1200
1400 0
200
400
600
800
1000
1200
1400
0
3
6
9
12
Average Vehicular Density(vehicles / 625m
2)
X (m)
Y (m)
Average Vehicular Density(vehicles / 625m
2)
Figure 6. Vehicular density: IDM
continuous perturbations caused by flows interaction at inter-
sections represents, by its nature, an unstable system. Thus,
the vehicular density and speed distributions showed next are
not representative of a steady state behavior, but rather give a
view on which is the general car mobility under the different
models.
The mobility models parameters used in these experiments
are listed in Table 1. We stress that different values of IDM-IM
k and IDM-LC p did not lead to significant differences in the
results, and that the IDM parameters were set to suitable real
world values.
In the following, we also report results obtained with the
Random Waypoint Model (RWP), in order to provide a bench-
mark of this popular model, which causes nodes to move with
random constant speed over a straight trajectory towards a des-
tination casually selected in the square area, and then to pause
for a random amount of time. Due to its nature, this model is
not bound by road constraints.
In Figure 4, the trend of the average speed versus the num-
ber of vehicles is shown. RWP and CSM, ignoring car-to-car
interactions, are not affected by the number of vehicles present
on the topology , leading to an unrealistically constant mean
speed. The mean velocity recorded with CSM is slightly lower
than that measured with RWP, even if the mean pause time is
shorter in CSM than in RWP. The reason is that CSM limits
nodes movement to the road topology, with pauses at every in-
tersection encountered on the path. Thus, the average distance
between subsequent pauses is reduced in CSM, with the con-
sequence of a lower average speed.
Proceedings of the 40th Annual Simulation Symposium (ANSS'07)0-7695-2814-7/07 $20.00 © 2007
Authorized licensed use limited to: IEEE Xplore. Downloaded on March 25, 2009 at 05:53 from IEEE Xplore. Restrictions apply.
Model RWP CSM FTM
Parameter speed pause speed pause smin smax κjam
Value unif [10, 20]m/s unif [0, 60]s unif [10, 20]m/s unif [0, 45]s 3m/s 20m/s 0.125car/m
Model IDM IDM-IM IDM-LC
Parameter v0 s0 T a b κ abias p athr
Value unif [10, 20]m/s 1m 0.5s 0.6m/s2 0.9m/s2 5 0.2m/s2 0.5 0.2m/s2
Table 1. Parameters value for the micro-mobility models
From Figure 4, modeling the vehicular mobility with FTM
produces a very high average speed, due to the fact that ve-
hicles never stop with this model, as the zero speed condi-
tion would cause a deadlock in the FTM formula. Probably,
a smaller value of the κjam parameter would have reduced this
effect, producing a lower and more realistic figure of the av-
erage velocity. However, the settings we chose force vehicles
to move at a minimum speed of 10 km/h when they are at a
distance of 3 m or less from each other, which represents a re-
alistic condition. As expected, FTM reproduces the average
speed reduction caused by the vehicular density growth, since
the increase of the number of cars traveling concurrently on
the same road reduces the fluid speed. However, the vehicu-
lar density distribution depicted in Figure 5 demonstrates the
non sufficient realism of this model. This distribution plot,
as well as the equivalent ones for the other mobility models
in the remainder of this Section, refers to the 30 vehicle/km
case. In the considered scenario, a high density is experienced
by the central segment marked as AB in Figure 3, which is
shared by many of the possible paths drivers can choose from.
The high quantity of cars driving through determines a reduc-
tion of the speed according to the model and creates an even
higher vehicular density, which is consistent with what would
happen in a real world situation. However, FTM reasons on a
per-edge basis and produces a constant car density over each
street, which results in the absence of traffic correlation over
connected roads. In our case, it can be noticed that the high car
density in AB suddenly disappear in roads after intersections
A and B (see Figure 3 for the mapping of letters to intersec-
tions). Moreover, as FTM ignores intersections, the average
number of vehicles at crossroads does not differ from that of
vehicles on roads nearby, which, again, is far from reality.
As far as IDM is concerned, the average speed curve in Fig-
ure 4 shows lower values when compared with that obtained
with FTM, and, quite surprisingly, appears to be affected by
the number of cars present on the topology. The speed re-
duction with respect to FTM is imputable to a more realistic
car-to-car interaction, which leads to queuing of fast vehicles
behind slow cars. The dependence from vehicular density has
instead a two-fold nature: first, the higher density increases
the probability of encountering slow vehicles, which generates
queues and forces a reduction on other drivers’ speed. Second,
there exists a side effect of the CanuMobiSim implementation,
200 400
600 800
1000 1200
1400 0
200
400
600
800
1000
1200
1400
0
3
6
9
12
Average Vehicular Density(vehicles / 625m
2)
X (m)
Y (m)
Average Vehicular Density(vehicles / 625m
2)
Figure 7. Vehicular density: IDM-IM stops
200 400
600 800
1000 1200
1400 0
200
400
600
800
1000
1200
1400
0
3
6
9
12
Average Vehicular Density(vehicles / 625m
2)
X (m)
Y (m)
Average Vehicular Density(vehicles / 625m
2)
Figure 8. Vehicular density: IDM-IM lights
which occurs when vehicles coming from different directions
and overlapping at intersections suddenly notice that the safety
distance condition is violated. According to the current imple-
mentation, they stop and wait for a distance s0 to be restored
before leaving the junction. Such a circumstance causes the av-
erage speed to decrease, and occurs more and more frequently
as the vehicular density grows. In Figure 6, the vehicular den-
sity proves that the realism of an accurate car-to-car interaction
model in urban scenarios is low, if intersection management is
not taken into account. Spikes at highly frequented intersec-
tions A, B and C are to impute to the implementation issue
explained above, while in general we can state that IDM does
not perform more realistically than FTM in an urban context.
Two different tests were run for IDM-IM, the first with in-
tersections regulated by stop signs, and the second with traffic
lights at road junctions. As observed in Figure 4, in the first
case the model produces a very low average speed, since cars
spend most of their time queued at intersections. The problem
is exacerbated as the density of vehicles increases and causes
Proceedings of the 40th Annual Simulation Symposium (ANSS'07)0-7695-2814-7/07 $20.00 © 2007
Authorized licensed use limited to: IEEE Xplore. Downloaded on March 25, 2009 at 05:53 from IEEE Xplore. Restrictions apply.
200 400
600 800
1000 1200
1400 0
200
400
600
800
1000
1200
1400
0
3
6
9
12
Average Vehicular Density(vehicles / 625m
2)
X (m)
Y (m)
Average Vehicular Density(vehicles / 625m
2)
Figure 9. Vehicular density: IDM-LC lights
longer queues. This can also be noticed by looking at the ve-
hicular density in Figure 7, where high vehicular densities, ac-
counting for long queues, are recorded in the neighborhoods
of the main intersections A, B, C and D. A realistic effect
of smooth vehicular density, increasing towards the congested
crossroads, is obtained with this model. It can be noticed that
such effect in not limited to single segments as it happened
with FTM, but also impacts adjacent roads.
When traffic lights with a period of 90 s are used to regu-
late traffic at intersections, vehicular mobility is improved with
respect to the stop sign case, especially in dense scenarios, as
proved by Figure 4. This could be expected, as traffic lights re-
place the slow “taking-turns” crossroads management induced
by stop signs with a faster “burst” mechanism, in which groups
of cars are allowed to cross the junction one after the other, thus
saving on acceleration delay. However, the mean speed is still
reduced when more cars are introduced in the road topology,
for the same reason observed in the stop sign case. An interest-
ing effect can be observed when the vehicular density is low,
as the stop sign case outperforms the traffic light one. This
occurs because, when the number of cars is small, the prob-
ability that a crossroad is free is high, thus passage is often
granted immediately with a stop handling of intersections, at
the cost of slowing and accelerating again. On the other hand,
when a traffic light management is considered, vehicles still
have to stop in presence of red traffic lights, even if there are
no other cars waiting to cross the intersection, and wait for the
light to turn green. The vehicular density, presented in Fig-
ure 8, appears consistent with the speed figure, as queuing at
highly visited intersections is still present, but noticeably re-
duced with respect to the previous IDM-IM scenario. Thanks
to the improved distribution of traffic over the whole topology,
the queuing phenomenon can now be observed at minor inter-
sections, where vehicles have to wait for green traffic lights.
Finally, we report the results obtained when IDM-LC is
employed as micro-mobility model. We considered two per-
direction lanes on each road, and traffic lights at intersections.
From Figure 4, modeling vehicular micro-mobility with IDM-
LC seems to avoid most of the speed decay effect discussed
before. This is an interesting result, motivated by the fact that
i) vehicles actually employ overtakings to avoid slow cars and
0
5
10
15
20
25
Road (m)
Tim
e (s
)
0 100 200 300 400 500 600 700 800
1800
2000
2200
2400
2600
2800
Figure 10. Vehicular speed shock waves
congested lanes, thus increasing the average velocity, and ii)
the presence of multiple lanes helps vehicular mobility in pres-
ence of densely populated intersections, as multiple cars can
pass through the intersection at the same time and reduce the
bottleneck effect of road junctions. In other words, the avail-
ability of two parallel unidirectional lanes on each road not
only physically doubles the capacity of the urban infrastruc-
ture, leading to a halved perceived vehicular density, but also
brings important correlated effects. In our case, the maxi-
mum simulated density of 50 vehicles/km would appear, for
the reasons explained before, as a density of less than 25 ve-
hicles/km, a condition which does not seem to generate severe
traffic congestion. The vehicular density measured with IDM-
LC is depicted in Figure 9 and shows that queuing phenomena
at intersections are almost equally distributed over the whole
topology. Minor intersections experience a higher density with
respect to the IDM-IM case as, in absence of critical conges-
tion situations at main junctions, vehicles are more uniformly
spread and their presence at smaller crossroads is more note-
worthy.
In a different test, we exploited the vehicular mobility de-
scription provided by VanetMobiSim to recreate a typical ef-
fect of vehicular traffic. In Figure 10, the shock waves pro-
duced on vehicular speed by a periodic perturbation are shown.
This result has been obtained with IDM-LC on a 1 km long,
unidirectional, double lane, straight road. Cars move towards
positive abscissae and a traffic light, located halfway and with
a period of 360 s, is used as the perturbation source. We can
notice that the red traffic light inhibits the movement of vehi-
cles, causing them to stop at 500 m. As more vehicles approach
the traffic light, a queue is formed, as shown by the decreasing
vehicular speed, but, when the traffic light turns green, queued
vehicles start flowing towards and through the second half of
the road. It is possible to see that the low speed shock wave
propagates in the opposite direction with respect to movement
of cars as time goes on. Shock waves are a common phenom-
ena of real world traffic. When long queues form in proximity
of perturbation sources (crowded intersections, toll stations, in-
flow ramps, etc.) the finite reaction time of drivers determines a
Proceedings of the 40th Annual Simulation Symposium (ANSS'07)0-7695-2814-7/07 $20.00 © 2007
Authorized licensed use limited to: IEEE Xplore. Downloaded on March 25, 2009 at 05:53 from IEEE Xplore. Restrictions apply.
Figure 11. Simulated vehicular mobility in the
Westwood area
delay in the propagation of movement. Thus, vehicles queued
far from the perturbation origin experience changes in veloc-
ity or local traffic density only a long time after the original
mobility change occurs at the perturbation.
As a further addition to the validation of the mobility gen-
erated by VanetMobiSim, Figure 11 shows a snapshot of the
vehicular mobility obtained with VanetMobiSim on the urban
area of Westwood in Los Angeles is overlap to a real map of the
same city section. The snapshot refers to a simulation involv-
ing IDM-IM, traffic lights at intersections, a random speed-
based path selection. Drivers thus take into account the path
length and the allowed speed along the path, making detours
if a path appears globally faster. The consequence can be seen
in Figure 11, where Wilshire Boulevard attracts the majority of
drivers, hoping to save time by using a large East-West com-
muting corridor instead of parallel streets. When the local ve-
hicular density exceeds the traffic lights management capabil-
ity, the traffic cluster pours out and cars start stacking up on the
surrounding streets and not only at the road junctions. These
congestion phenomena can be easily observed in real-life situ-
ations.
4. Conclusions and Future Work
In this paper we presented VanetMobiSim, an extension to
the CanuMobiSim user mobility framework capable of produc-
ing realistic vehicular mobility traces for several network sim-
ulators. We reviewed the macroscopic and microscopic mobil-
ity descriptions of CanuMobiSim, and detailed the additions
to both scopes brought by VanetMobiSim. Simulation results
were presented and discussed, trying to understand the differ-
ences between various micro-mobility models, in terms of ve-
hicular density and speed distribution.
By taking a comprehensive look at the results obtained, it
appears clear that the detail level of the micro-mobility models
implemented by CanuMobiSim is not sufficient to reproduce
realistic vehicular traffic traces. The increased degree of de-
tail introduced by the micro-mobility models of the VanetMo-
biSim extension, and the possibility of their interaction with the
new macro-mobility description appear necessary to reproduce
real world phenomena. In particular, the progressive introduc-
tion of stops signs, traffic lights, multiple lanes and overtakings
demonstrates how the modeling of each of these features brings
noticeable changes to the system performance.
From a networking point of view, the differences observed
between different micro-mobility models, in terms of vehicles
and speed distribution, queuing dynamics and presence and
size of clusters may heavily affect the connectivity of VANETs,
and, consequently, the performance of ad-hoc network proto-
cols. It is part of future work to investigate the actual impact
of these different traffic phenomena on a vehicular network, so
to understand which factors must be considered and which can
be neglected for a confident VANETs simulation study. Also, a
very important factor when simulating highly mobile networks
is the radio propagation model. Results obtained without ac-
counting for the impact of large obstacles, such as buildings,
on the radio signal propagation can hardly be realistic. We are
thus interested in studying this aspect, taking benefit from the
availability of a detailed topology description to introduce a
new component in VanetMobiSim, capable of generating radio
propagation information for network simulators.
5 References
[1] J. Harri, F.Filali, and C. Bonnet, ”Mobility Models for Vehicular
Ad Hoc Networks: A Survey and Taxonomy”, Technical Report
RR-06-168, Institut Eurecom, January 2007.
[2] J. Harri, M. Fiore, F. Filali, and C. Bonnet, ”A Realistic Mobility
Simulator for Vehicular Ad Hoc Networks”, Technical Report RR-
05-150, Institut Eurecom, January 2007.
[3] VanetMobiSim, http://vanet.eurecom.fr.
[4] CANU, http://canu.informatik.uni-stuttgart.de.
[5] ns-2 http://www.isi.edu/nsnam/ns.
[6] QualNet, http://www.scalable-networks.com
[7] GloMoSim, http://pcl.cs.ucla.edu/projects/glomosim.
[8] Ertico, http://www.ertico.com.
[9] U.S. Census Bureau TIGER system database,
http://www.census.gov/geo/www/tiger.
[10] A. Jardosh, E. Belding-Royer, K. Almeroth, and S. Suri, ”Toward
realistic mobility models for mobile ad hoc networks”, MobiCom
2003, San Diego, CA.
[11] J. Tian, J. Haehner, C. Becker, I. Stepanov, K. Rothermel ”Graph-
Based Mobility Model for Mobile Ad-Hoc Network Simulation”,
Annual Simulation Symposium 2002.
[12] C. Bettstetter, ”Smooth is Better than Sharp: A Random Mobil-
ity Model for Simulation of Wireless Networks”, MSWiM 2001,
Rome, Italy.
[13] I. Seskar, S. Marie, J. Holtzman, J. Wasserman, ”Rate of Location
Area Updates in Cellular Systems”, VTC 1992, Denver, CO.
[14] M. Trieber, A. Hennecke, D. Helbing, ”Congested traffic states in
empirical observations and microscopic simulations”, Phys. Rev. E
62, Issue 2, August 2000.
[15] M. Treiber, D. Helbing, ”Realistische Mikrosimulation von
Strassenverkehr mit einem einfachen Modell”, Symposium Sim-
ulationstechnik ASIM 2002, Rostock, Germany.
Proceedings of the 40th Annual Simulation Symposium (ANSS'07)0-7695-2814-7/07 $20.00 © 2007
Authorized licensed use limited to: IEEE Xplore. Downloaded on March 25, 2009 at 05:53 from IEEE Xplore. Restrictions apply.